import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import plotly.express as px
import plotly.io as pio
pio.renderers.default = 'notebook'

df=pd.read_csv('Cleaned_Clothing.csv')

df.head()

	Customer ID	Customer Name	Product	Product Category	Size	Color	Cost (in NPR)	Quantity	Total Cost (in NPR)	Purchase Date	Store Location	Purchase Method	Foot Traffic	Gender	Product Rating	Customer Type
0	CUST001	Customer_1	Dress	Casualwear	L	Black	1800	4	7200	2024-04-30	Bhaktapur	Visited Store	132	Male	2	New
1	CUST002	Customer_2	Shirt	Casualwear	L	Blue	1737	2	3474	2023-12-17	Lalitpur	Website	471	Female	2	New
2	CUST003	Customer_3	Skirt	Bottomwear	S	Black	648	3	1944	2023-11-19	Boudha	Visited Store	488	Male	4	Returning
3	CUST004	Customer_4	Blazer	Bottomwear	XL	Yellow	2603	1	2603	2024-04-30	Thamel	Visited Store	309	Male	2	Returning
4	CUST005	Customer_5	T-shirt	Outerwear	M	Blue	780	1	780	2024-09-16	Bhaktapur	Visited Store	395	Male	2	Returning

df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 100 entries, 0 to 99
Data columns (total 16 columns):
 #   Column               Non-Null Count  Dtype 
---  ------               --------------  ----- 
 0   Customer ID          100 non-null    object
 1   Customer Name        100 non-null    object
 2   Product              100 non-null    object
 3   Product Category     100 non-null    object
 4   Size                 100 non-null    object
 5   Color                100 non-null    object
 6   Cost (in NPR)        100 non-null    int64 
 7   Quantity             100 non-null    int64 
 8   Total Cost (in NPR)  100 non-null    int64 
 9   Purchase Date        100 non-null    object
 10  Store Location       100 non-null    object
 11  Purchase Method      100 non-null    object
 12  Foot Traffic         100 non-null    int64 
 13  Gender               100 non-null    object
 14  Product Rating       100 non-null    int64 
 15  Customer Type        100 non-null    object
dtypes: int64(5), object(11)
memory usage: 12.6+ KB

# Sorting
df = df.sort_values("Customer Name")
df.head()

	Customer ID	Customer Name	Product	Product Category	Size	Color	Cost (in NPR)	Quantity	Total Cost (in NPR)	Purchase Date	Store Location	Purchase Method	Foot Traffic	Gender	Product Rating	Customer Type
0	CUST001	Customer_1	Dress	Casualwear	L	Black	1800	4	7200	2024-04-30	Bhaktapur	Visited Store	132	Male	2	New
9	CUST010	Customer_10	Shorts	Bottomwear	S	Red	623	1	623	2024-05-13	Kathmandu	Visited Store	144	Female	5	Returning
99	CUST100	Customer_100	T-shirt	Topwear	M	Red	1287	5	6435	2024-01-13	Bhaktapur	Visited Store	150	Male	5	New
10	CUST011	Customer_11	Blazer	Formalwear	M	Yellow	2142	2	4284	2024-09-29	Thamel	Visited Store	418	Male	2	Returning
11	CUST012	Customer_12	T-shirt	Topwear	L	Pink	620	1	620	2024-02-21	Thamel	Visited Store	127	Male	5	New

df.isnull().sum()

Customer ID            0
Customer Name          0
Product                0
Product Category       0
Size                   0
Color                  0
Cost (in NPR)          0
Quantity               0
Total Cost (in NPR)    0
Purchase Date          0
Store Location         0
Purchase Method        0
Foot Traffic           0
Gender                 0
Product Rating         0
Customer Type          0
dtype: int64

# Dropping Unnecessary Columns
df = df.drop(["Customer ID", "Foot Traffic"], axis=1)
df.head()

	Customer Name	Product	Product Category	Size	Color	Cost (in NPR)	Quantity	Total Cost (in NPR)	Purchase Date	Store Location	Purchase Method	Gender	Product Rating	Customer Type
0	Customer_1	Dress	Casualwear	L	Black	1800	4	7200	2024-04-30	Bhaktapur	Visited Store	Male	2	New
9	Customer_10	Shorts	Bottomwear	S	Red	623	1	623	2024-05-13	Kathmandu	Visited Store	Female	5	Returning
99	Customer_100	T-shirt	Topwear	M	Red	1287	5	6435	2024-01-13	Bhaktapur	Visited Store	Male	5	New
10	Customer_11	Blazer	Formalwear	M	Yellow	2142	2	4284	2024-09-29	Thamel	Visited Store	Male	2	Returning
11	Customer_12	T-shirt	Topwear	L	Pink	620	1	620	2024-02-21	Thamel	Visited Store	Male	5	New

# Rename the column in the DataFrame
df.rename(columns={'Cost (in NPR)': 'Cost'}, inplace=True)
df.rename(columns={'Total Cost (in NPR)': 'Total_Cost'}, inplace=True)

#mapping visited store  as 1 and website for 0
purchase_method_mapping= {"Visited Store": 1, "Website": 0}
df["Purchase Method"] = df["Purchase Method"].map(purchase_method_mapping)

df.head()

	Customer Name	Product	Product Category	Size	Color	Cost	Quantity	Total_Cost	Purchase Date	Store Location	Purchase Method	Gender	Product Rating	Customer Type
0	Customer_1	Dress	Casualwear	L	Black	1800	4	7200	2024-04-30	Bhaktapur	1	Male	2	New
9	Customer_10	Shorts	Bottomwear	S	Red	623	1	623	2024-05-13	Kathmandu	1	Female	5	Returning
99	Customer_100	T-shirt	Topwear	M	Red	1287	5	6435	2024-01-13	Bhaktapur	1	Male	5	New
10	Customer_11	Blazer	Formalwear	M	Yellow	2142	2	4284	2024-09-29	Thamel	1	Male	2	Returning
11	Customer_12	T-shirt	Topwear	L	Pink	620	1	620	2024-02-21	Thamel	1	Male	5	New

#mapping customer type as 1 for new and 0 for existing
customer_type_mapping= {
    "New": 1, 
    "Returning": 0
    }
df["Customer Type"] = df["Customer Type"].map(customer_type_mapping)
df.head()

	Customer Name	Product	Product Category	Size	Color	Cost	Quantity	Total_Cost	Purchase Date	Store Location	Purchase Method	Gender	Product Rating	Customer Type
0	Customer_1	Dress	Casualwear	L	Black	1800	4	7200	2024-04-30	Bhaktapur	1	Male	2	1
9	Customer_10	Shorts	Bottomwear	S	Red	623	1	623	2024-05-13	Kathmandu	1	Female	5	0
99	Customer_100	T-shirt	Topwear	M	Red	1287	5	6435	2024-01-13	Bhaktapur	1	Male	5	1
10	Customer_11	Blazer	Formalwear	M	Yellow	2142	2	4284	2024-09-29	Thamel	1	Male	2	0
11	Customer_12	T-shirt	Topwear	L	Pink	620	1	620	2024-02-21	Thamel	1	Male	5	1

#mapping male as 1 and female as 0
gender_mapping= {
    'Male': 1,
    'Female': 0
    }
df["Gender"] = df["Gender"].map(gender_mapping)

df.info()

<class 'pandas.core.frame.DataFrame'>
Index: 100 entries, 0 to 98
Data columns (total 14 columns):
 #   Column            Non-Null Count  Dtype 
---  ------            --------------  ----- 
 0   Customer Name     100 non-null    object
 1   Product           100 non-null    object
 2   Product Category  100 non-null    object
 3   Size              100 non-null    object
 4   Color             100 non-null    object
 5   Cost              100 non-null    int64 
 6   Quantity          100 non-null    int64 
 7   Total_Cost        100 non-null    int64 
 8   Purchase Date     100 non-null    object
 9   Store Location    100 non-null    object
 10  Purchase Method   100 non-null    int64 
 11  Gender            100 non-null    int64 
 12  Product Rating    100 non-null    int64 
 13  Customer Type     100 non-null    int64 
dtypes: int64(7), object(7)
memory usage: 11.7+ KB

#mapping the size
size_mapping= {
    'S': 1,
    'M': 2,
    'L': 3,
    'XL': 4
    }
df["Size"] = df["Size"].map(size_mapping)

# Ratings Category (Low, Medium, High)
def categorize_rating(rating):
    if rating <= 2.5:
        return 'Low'
    elif rating <= 3.5:
        return 'Medium'
    else:
        return 'High'

df['Rating_Categoy'] = df['Product Rating'].apply(categorize_rating)
df.head()

	Customer Name	Product	Product Category	Size	Color	Cost	Quantity	Total_Cost	Purchase Date	Store Location	Purchase Method	Gender	Product Rating	Customer Type	Rating_Categoy
0	Customer_1	Dress	Casualwear	3	Black	1800	4	7200	2024-04-30	Bhaktapur	1	1	2	1	Low
9	Customer_10	Shorts	Bottomwear	1	Red	623	1	623	2024-05-13	Kathmandu	1	0	5	0	High
99	Customer_100	T-shirt	Topwear	2	Red	1287	5	6435	2024-01-13	Bhaktapur	1	1	5	1	High
10	Customer_11	Blazer	Formalwear	2	Yellow	2142	2	4284	2024-09-29	Thamel	1	1	2	0	Low
11	Customer_12	T-shirt	Topwear	3	Pink	620	1	620	2024-02-21	Thamel	1	1	5	1	High

df['Product Category'].unique()

array(['Casualwear', 'Bottomwear', 'Topwear', 'Formalwear', 'Outerwear'],
      dtype=object)

#mapping product wear
product_category_mapping= {
    'Casualwear': 1,
    'Formalwear': 2,
    'Bottomwear': 3,
    'Outerwear': 4,
    'Topwear': 5
    }
df["Product Category"] = df["Product Category"].map(product_category_mapping)

Visualizations

#visualization based on the store location and the purchase method
import plotly.express as px

df['store'] = df['Store Location'].astype('category')
fig = px.scatter(df, x='store', y='Cost', color='Purchase Method', 
                 title='Scatter Plot of Cost by Store Location and Purchase Method')
fig.show()

# Group by Gender and calculate the average Cost
grouped_df = df.groupby('Gender')['Cost'].mean().reset_index()

# Create a bar plot with specific colors for male and female
fig = px.bar(grouped_df, x='Gender', y='Cost', title='Average Cost by Gender',
             color='Gender', color_discrete_map={'Male': 'blue', 'Female': 'pink'})
fig.show()

# Group by Store Location and sum the Total Cost
grouped_df = df.groupby('Store Location')['Total_Cost'].sum().reset_index()

# Create a bar plot
fig = px.bar(grouped_df, x='Store Location', y='Total_Cost', title='Total Cost by Store Location')
fig.show()

# Group by Purchase Method and sum the Quantity
grouped_df = df.groupby('Purchase Method')['Quantity'].sum().reset_index()
fig = px.pie(grouped_df, names='Purchase Method', values='Quantity', title='Total Quantity by Purchase Method', hole=0.3)
fig.show()

# Group by Product Category and Size, and sum the Total Cost
grouped_df = df.groupby(['Product Category', 'Size'])['Total_Cost'].sum().reset_index()

# Create a line plot
fig = px.line(grouped_df, x='Product Category', y='Total_Cost', color='Size', 
              title='Total Cost by Product Category and Size', markers=True)
fig.show()

# Group by Product Category and calculate the average Product Rating
grouped_df = df.groupby('Product Category')['Product Rating'].mean().reset_index()
fig = px.scatter(grouped_df, x='Product Category', y='Product Rating', title='Average Product Rating by Product Category',
                 size='Product Rating', color='Product Category')
fig.show()

# Group by Store Location and sum the Total Cost
grouped_df = df.groupby('Store Location')['Total_Cost'].sum().reset_index()
fig = px.bar(grouped_df, x='Store Location', y='Total_Cost', title='Total Cost by Store Location',color='Store Location')
fig.show()

# Group by Gender and Purchase Method, and count the occurrences
grouped_df = df.groupby(['Gender', 'Purchase Method']).size().reset_index(name='Count')
fig = px.pie(grouped_df, names='Purchase Method', values='Count', 
             title='Gender-wise Purchase Method Distribution', facet_col='Gender')
fig.show()

# Group by Product Category and calculate the average Product Rating
grouped_df = df.groupby('Product Category')['Product Rating'].mean().reset_index()
fig = px.bar(grouped_df, x='Product Category', y='Product Rating', title='Average Product Rating by Product Category',
             color='Product Category')
fig.show()

df.columns

Index(['Customer Name', 'Product', 'Product Category', 'Size', 'Color', 'Cost',
       'Quantity', 'Total_Cost', 'Purchase Date', 'Store Location',
       'Purchase Method', 'Gender', 'Product Rating', 'Customer Type',
       'Rating_Categoy', 'store'],
      dtype='object')

df.describe()

	Product Category	Size	Cost	Quantity	Total_Cost	Purchase Method	Gender	Product Rating	Customer Type
count	100.00000	100.000000	100.000000	100.000000	100.00000	100.000000	100.000000	100.00000	100.000000
mean	3.15000	2.510000	2634.640000	3.230000	8592.31000	0.510000	0.440000	2.99000	0.480000
std	1.34371	1.029808	1383.482113	1.462356	6176.86699	0.502418	0.498888	1.46677	0.502117
min	1.00000	1.000000	533.000000	1.000000	620.00000	0.000000	0.000000	1.00000	0.000000
25%	2.00000	2.000000	1326.750000	2.000000	3452.25000	0.000000	0.000000	2.00000	0.000000
50%	3.00000	3.000000	2550.500000	3.000000	7446.00000	1.000000	0.000000	3.00000	0.000000
75%	4.00000	3.000000	3888.500000	5.000000	12371.25000	1.000000	1.000000	4.00000	1.000000
max	5.00000	4.000000	4979.000000	5.000000	24695.00000	1.000000	1.000000	5.00000	1.000000

corr = df.corr(numeric_only=1)
corr

	Product Category	Size	Cost	Quantity	Total_Cost	Purchase Method	Gender	Product Rating	Customer Type
Product Category	1.000000	-0.019344	0.047035	-0.063999	0.044899	-0.009725	0.036163	0.149395	0.251515
Size	-0.019344	1.000000	0.197383	-0.018311	0.153153	-0.117332	-0.126617	-0.103585	-0.028911
Cost	0.047035	0.197383	1.000000	0.041151	0.722298	0.091731	-0.112076	-0.231336	0.065437
Quantity	-0.063999	-0.018311	0.041151	1.000000	0.639537	-0.161267	-0.070889	0.076431	0.013206
Total_Cost	0.044899	0.153153	0.722298	0.639537	1.000000	-0.001962	-0.160343	-0.082812	0.052109
Purchase Method	-0.009725	-0.117332	0.091731	-0.161267	-0.001962	1.000000	0.143465	-0.061544	-0.059259
Gender	0.036163	-0.126617	-0.112076	-0.070889	-0.160343	0.143465	1.000000	-0.021534	-0.125809
Product Rating	0.149395	-0.103585	-0.231336	0.076431	-0.082812	-0.061544	-0.021534	1.000000	-0.075707
Customer Type	0.251515	-0.028911	0.065437	0.013206	0.052109	-0.059259	-0.125809	-0.075707	1.000000

# Heatmap
plt.figure(figsize=(12, 8))
sns.heatmap(corr, annot=True, fmt='.2f', cmap='magma', linewidths=0.5)
plt.title('Correlation Heatmap')
plt.show()

model Training

Features = df[['Total_Cost','Purchase Method','Product Category','Size','Quantity']]
Target = df['Cost']

#splitting the data
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(Features, Target, test_size=0.2, random_state=0)

# Gradient Boosting Regressor
from sklearn.ensemble import GradientBoostingRegressor
from sklearn.metrics import mean_absolute_error, r2_score

gbr = GradientBoostingRegressor()
# Train the model
gbr.fit(X_train, y_train)
y_pred = gbr.predict(X_test)

# Model Evaluation
mae = mean_absolute_error(y_test, y_pred)
r2 = r2_score(y_test, y_pred)
print("Mean Absolute Error: ", mae)
print("R2 Score: ", r2)

Mean Absolute Error:  200.90494036792384
R2 Score:  0.9729606885493439

# Visualization of predicted vs actual costs
fig_cost = px.scatter(x=y_test, y=y_pred, labels={'x': 'Actual Cost', 'y': 'Predicted Cost'}, title='Actual vs Predicted Cost')
fig_cost.add_shape(
    type="line", line=dict(dash='dash'),
    x0=y_test.min(), y0=y_test.min(),
    x1=y_test.max(), y1=y_test.max()
)
fig_cost.update_layout(paper_bgcolor="white")
fig_cost.show()

cost_list = [[predicted] for actual, predicted in zip(y_test, y_pred)]

from sklearn.ensemble import GradientBoostingRegressor
from sklearn.metrics import r2_score, mean_absolute_error
from sklearn.model_selection import cross_val_score

gbr = GradientBoostingRegressor(n_estimators=300, learning_rate=0.2, max_depth=3, random_state=42,loss='squared_error')
gbr.fit(X_train, y_train)

y_pred = gbr.predict(X_test)

cv_scores_r2 = cross_val_score(gbr, Features, Target, cv=5, scoring='r2')
print("Cross-validation R2 scores for Gradient Boosting Regressor:", cv_scores_r2)
print("Mean R2 score:", cv_scores_r2.mean())

# Perform cross-validation for Mean Absolute Error
cv_scores_mae = cross_val_score(gbr, Features, Target, cv=5, scoring='neg_mean_absolute_error')
cv_scores_mae = -cv_scores_mae  # Convert to positive values
print("Cross-validation MAE scores for Gradient Boosting Regressor:", cv_scores_mae)
print("Mean MAE score:", cv_scores_mae.mean())

Cross-validation R2 scores for Gradient Boosting Regressor: [0.97341073 0.96872318 0.9516513  0.94338487 0.95744448]
Mean R2 score: 0.9589229130407503
Cross-validation MAE scores for Gradient Boosting Regressor: [180.71856597 207.74159584 229.73779601 201.30292391 215.83694738]
Mean MAE score: 207.06756582212523

Random Forest Regressor

from sklearn.ensemble import RandomForestRegressor
# basic model
rfr = RandomForestRegressor()
rfr.fit(X_train, y_train)
y_pred = rfr.predict(X_test)

mae = mean_absolute_error(y_test, y_pred)
r2 = r2_score(y_test, y_pred)
print("Mean Absolute Error: ", mae)
print("R2 Score: ", r2)

Mean Absolute Error:  419.21450000000004
R2 Score:  0.8840462075478562

#plot tree diagram based model
from sklearn.tree import plot_tree
plt.figure(figsize=(20, 10))
plot_tree(rfr.estimators_[0], filled=True, feature_names=Features.columns)
plt.show()

Hyperparameter and cross validation

#hyperparameter tuning for random forest regressor
rfr_params = {
    'n_estimators': 300,
    'max_depth': 3,
    'random_state': 42
}
rfr = RandomForestRegressor(**rfr_params)

# Fit the model
rfr.fit(X_train, y_train)

# Predict on the test set
y_pred_reg = rfr.predict(X_test)

# Perform cross-validation for R2 score
cv_scores_r2 = cross_val_score(rfr, Features, Target, cv=5, scoring='r2')
print("Cross-validation R2 scores for Random Forest Regressor:", cv_scores_r2)
print("Mean R2 score:", cv_scores_r2.mean())

# Perform cross-validation for Mean Absolute Error
cv_scores_mae = cross_val_score(rfr, Features, Target, cv=5, scoring='neg_mean_absolute_error')
cv_scores_mae = -cv_scores_mae 
print("Cross-validation MAE scores for Random Forest Regressor:", cv_scores_mae)
print("Mean MAE score:", cv_scores_mae.mean())

Cross-validation R2 scores for Random Forest Regressor: [0.85507448 0.80738194 0.81554827 0.7633761  0.77008654]
Mean R2 score: 0.8022934659816918
Cross-validation MAE scores for Random Forest Regressor: [403.70004968 564.66271815 468.18076216 559.51067536 584.78026227]
Mean MAE score: 516.1668935224409

# Visualization of predicted vs actual costs after using Random Forest Regressor
fig_cost = px.scatter(x=y_test, y=y_pred_reg, labels={'x': 'Actual Cost', 'y': 'Predicted Cost'}, title='Actual vs Predicted Cost')
fig_cost.add_shape(
    type="line",
    x0=y_test.min(),
    y0=y_test.min(),
    x1=y_test.max(),
    y1=y_test.max(),
    line=dict(color="Red"),
    )
fig_cost.update_layout(paper_bgcolor="white")
fig_cost.show()

import plotly.graph_objs as go
from plotly.offline import init_notebook_mode, iplot

init_notebook_mode(connected=True)
fig = go.Figure(data=[go.Scatter(x=[1, 2, 3], y=[4, 5, 6])])
iplot(fig)